Twice Upon a Time: Complementary Representations for Time-Series Reasoning

Key points

• We instead introduce complementary representations as a design principle for time-series reasoning.1
• Our model combines two visual views: line plots preserve absolute measurements and the observed coordinate frame, while normalized delay embeddings factor out positive amplitude rescaling and offsets while exposing temporal structure.1
• Dedicated visual encoders integrate both views into a shared language model through visual tokens and multi-level feature fusion.1
• Together, these results demonstrate that complementary representations provide an effective design principle for general-purpose TSLMs, enabling models to exploit temporal structure without sacrificing access to the original measurement frame.1

Sources (1)

• [1]Twice Upon a Time: Complementary Representations for Time-Series Reasoning
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 05:52 PM

  Time Series Language Models (TSLMs) enable language-based understanding and reasoning over temporal data, yet a fundamental question remains: what information should their time-series representations preserve, and what variation should they factor out?

  We instead introduce complementary representations as a design principle for time-series reasoning.

Extractive summary: sentences quoted from the sources.